Reliability of retrospective assessments of childhood experiences in Germany.

The retrospective assessment of adverse childhood experiences is a widely used technique. Few studies report heterogeneous results about the reliability and objectivity. A sample of 100 patients was independently interviewed by means of the Mainz Structured Biographical Interview by two different interviewers with mean time lag of 2.2 years. Reliability of reports of family situation, physical abuse, sexual abuse, and protective factors was assessed. The results show moderate to good reliability for most childhood experiences. Divorce/separation of parents had a kappa of .95, severe sexual abuse a kappa of .64, and regular harsh physical abuse a kappa of .56; protective factors were in the .50 range. The key features for obtaining good reliability seem to be the concreteness of the question and accuracy of the coding categories. There is some, but little, evidence that sensitive issues such as sexual abuse are more prone to errors in assessment than simple ones.